World's First Ethernet Cable

It's a bit like Bart Simpsons reply to the question 'Do you know long division?'
'I know of long division'
Except I'd be talking about Manchester encoding and stuff.

Del

OK, I'll start with Manchester encoding:

Wikipedia gives a better basic description than I could. But it probably misses the main points as regards the impact on Ethernet

The generally perceived problems that are solved by Manchester encoding both result from the fact that random data can contain long sequences of 1's or zeros. Such a sequence is indistinguishable from DC - so:
. transformer or capacitor coupling would mean that DC recovery would be required, and
. there can be periods where there is no timing data.

Less widely understood as an issue regarding loss in cable transmission:
. Loss in the cable increases with increasing frequency (approximately proportional to √f). This requires correction (equalisation) when the variation across the used frequency range becomes excessive. Manchester coding shifts the frequency range so that this variation is reduced. Consequently, for this particular case, Manchester encoding allows transmission distances can to be increased by more than a factor of 3 before equalisation is needed.

Applicability to Ethernet:
Ethernet establishes timing using a preamble, and the original data patterns included sufficient parity information to prevent loss of timing - so the loss of timing data need not have been an issue.
Similarly, DC restoration methods were well-known and understood, as was the complementary aspect of linear equalisation.
So, IMHO, the reason for using Manchester would have been simplicity of implementation in the lab. This is based on non-linear DC recovery and linear equalisation being already well-understood and [also IMHO] that equalisation would have allowed the use of cheaper cabling, with lower overall cost in volume than the high-end cables and connectors originally specified).

Ta